The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs automated validation by executing internal Python scripts plan-validate.py and plan-metrics.py located in a shared scripts directory. These tools are used to verify the structure and metrics of generated task lists.
[PROMPT_INJECTION]: The skill processes untrusted external content from prd.md and design.md files to generate implementation tasks. This creates a surface for indirect prompt injection where malicious instructions embedded in requirement documents could influence the agent's behavior or the resulting task list.
Ingestion points: Reads project requirements from plans/<name>/design.md and plans/<name>/prd.md, and reads existing task lists from plans/<name>/tasks.md for updates.
Boundary markers: The skill lacks explicit instructions or delimiters to isolate and ignore potentially malicious instructions embedded within the user-provided markdown documents.
Capability inventory: The skill can execute shell commands via Python subprocesses and perform file system write operations to plans/<name>/tasks.md.
Sanitization: No content sanitization or instruction filtering is performed on the input documents before they are analyzed for task decomposition.

design-to-tasks